Stereotypes and prejudices remain an all-too-common and harmful influence on healthy social interaction. Mental patients, expatients, ethnic groups, minorities, old people, and gender groups all can suffer when perceived in ways that are based on their group membership, rather than on their other individual characteristics. Perceiving individuals in terms of group membership can lead to inaccurate and unfair responses. Recent social psychological research on stereotyping and prejudice has taken a cognitive perspective. In this view, social-group categories function for the perceiver as do any other categories; they serve to classify and identify individuals, thereby saving time and effort. Thus, social categorization is believed to result from normal and natural cognitive processes. This well-taken viewpoint, however, neglects the role of motivations that facilitate or undermine category-based impression formation. Currently-funded work examines the role of several motivations: outcome-dependency on a peer, competition with a peer, authority influence, and altruistic motivations. The proposed research examines a potential guiding framework for further work on motivation and category-based processes. Kelley's theory of dyadic interdependence is integrated with the PI's model of an impression formation continuum that ranges from category-based to individualing processes. The resulting premises are simple but fruitful: First, people seek prediction and control over positive outcomes. Second, they attend to others who control those outcomes. Consequently, the structure of interdependence -- specifically, its symmetry, its degree of facilitation, and its strength -- will predict the direction and nature of a person's attention in a dyad. Third, attention allows, but does not guarantee, individualing impression formation. Nine specific hypotheses and 23 laboratory studies are derived from this framework. Together, they examine the effects of interdependence structures on impression formation within a dyad that involves members of two distinct social groups. Each study exemplifies a common circumstance in which cross-group dyads interact, and each tests a different feature of the framework. Together, they investigate a potential overrearching framework for understanding the effects of motivation on category-based responses to mental patients and other stigmatized groups.